In the present study, we attempted to select DRV-resistant HIV-1 variants by propagating a mixture of HIV-1 variants (HIV-1MIX) isolated from 8 patients with AIDS, who had received antiretroviral therapy over 32-83 months and were not responding antiretroviral regimens in the presence of DRV. DRV-resistant HIV-1 at passage 51 (HIV-1MIXP51) replicated in the presence of 5 microM DRV and contained 14 mutations (L10I, I15V, K20R, L24I, V32I, L33F, M36I, M46L, I54M, L63P, K70Q, V82I, I84V, and L89M) in the protease-encoding region. At first, we attempted to select DRV-resistant HIV-1 variants by propagating a wild type HIV-1 strain (HIV-1NL4-3) in the presence of DRV. However, even after 90 passages (24 months), DRV-resistant HIV-1 variants did not emerge and the virus only poorly replicated and failed to replicate in the presence of &gt;0.1 microM DRV. The results confirm DRV's high intrinsic barrier to resistance but suggest resistance could be a problem in PI-naive people who start DRV/rtv then get superinfected with PI-resistant virus. A set of mutations (V11I, V32I, L33F, I47V, I50V, I54L/M, G73S, L76V, I84V, and L89V) was identified in HIV-1 isolated from those failing DRV-containing regimens that were associated with a diminished virological response to DRV/r at week 24 in the POWER studies. The most common mutations identified were V32I, L33F, I54M/L, I84V, and L89V. In the present study, HIV-1 variants resistant to DRV, which replicated in the presence of 1 and 5 microM DRV, emerged by passages 39 and 51, respectively. The protease-encoding region of the proviral DNA isolated from infected MT-4 cells was cloned and sequenced at passages 1, 10, 30, and 51 upon DRV selection. On passage 1, the virus had L10I, I15V, K20R, L24I, M36I, M46L, F53S, I54V, I62V, L63P, K70Q, V82A, and L89M substitutions compared to wild-type HIV-1NL4-3. By passage 10 and beyond, the virus additionally acquired a V32I substitution. By passage 30 and beyond, the virus contained an I84V substitution. By passage 51, the virus had acquired L33F, I54M, and V82I substitutions, and was found to contain 14 mutations including L10I, I15V, K20R, L24I, V32I, L33F, M36I, M46L, I54M, L63P, K70Q, V82I, I84V, and L89M in the protease-encoding region. It is of note that the four mutations (V32I, L33F, I54M, and I84V) HIV-1 acquired in the present study were the ones identified in highly DRV-resistant HIV-1 variants. With respect to HIV-1's acquisition of cross-resistance to TPV and DRV, our recent results showed both compounds blocked the dimerization of HIV-1 protease in the FRET-based-HIV-1-expression assay. Since HIV-1MIXP51 can propagate in the presence of 5 microM DRV, it is likely that DRV is no longer capable of inhibiting the dimerization of the protease with the mutations seen in HIV-1MIXP51. Considering that conventional protease inhibitors such as SQV, RTV, NFV, APV, and LPV failed to block the dimerization of HIV-1 protease, it appears that the activity to inhibit the proteolytic function of protease is independent from that to inhibit the dimerization of HIV-1 protease, although it should be determined what mutations (a single mutation or combined mutations) seen in HIV-1MIXP51 are responsible for the viral acquisition of the ability to escape from DRV's protease dimerization inhibition. It is also to be determined whether what mutations seen in HIV-1MIXP51 can confer resistance to TPV on HIV-1. During reverse transcription, reverse transcriptase is known to frequently switch a template from one genomic RNA strand to another, yielding recombinant proviral DNA, which represents a mosaic consisting of multiple parent genomic components. Indeed, there is firm evidence that a single CD4+ target cell can be infected with multiple HIV-1 virions both in vitro and in vivo. If resultant proviral DNA acquires mutations in one strand that confer HIV-1 resistance to one drug and mutations in the other strand that are associated to HIV-1 resistance to the other drug, such proviral DNA-containing daughter virions will be resistant to both drugs, a process called homologous recombination. Homologous recombination thus highly likely accelerates the emergence of multi-drug- and multi-class-drug-resistance in infected individuals. In the present work, the nucleic acid sequence of the protease-encoding region of HIV-1MIXP51 was virtually identical to that of HIV-1CR, suggesting that HIV-1C had resistance to DRV, immediately predominated other 7 HIV-1 isolates, and continued to propagate in the presence of DRV, suggesting no involvement of homologous recombination in the emergence of highly DRV-resistant HIV-1 variants. Thus, we further conducted a selection experiment using HIV-1C as a starting HIV-1 isolate. However, HIV-1C's DRV resistance acquisition substantially delayed compared to DRV resistance acquisition of HIV-1MIXP51 in two independent selection experiments. Therefore, we determined the nucleic acid sequence of the gag gene in both HIV-1MIXP51 and HIV-1CR and readily found that the former had two mutations (H219Q and I223V) in the CypA-binding loop and I247V by passage 10, while the latter lacked all these three mutations. Among the 8 isolate used, only the original HIV-1G contained these three mutations. The H219Q substitution in the viral CypA binding loop, a polymorphic amino acid change, has been shown to confer replication advantage on HIV-1 in CypA-rich target cells. As expected, both HIV-1MIXP39 and HIV-1MIXP51 had substantially less content of CypA within virions. When we generated four different clones that contained a variety of AA substitutions identified in the protease- and gag-encoding genes of the DRV-selected mixture population, only rHIVp1 replicated while other recombinant HIV-1 clones failed to replicate, although all the amino acid substitutions identified in the protease- and gag-encoding genes of the replicative mixture HIV-1 population were introduced to such recombinant clones. When we generated recombinant HIV-1 virions with two AA substitutions mutated back (H219Q/I223V to H219/I223, designated as rHIVp530-MB), these virions also failed to replicate. Suspecting that other amino acid substitutions residing in a minor HIV-1 population helped the mixture population replicate through homologous recombination in the mixture population, we added such an amino acid substitution, A196T or A196S, to rHIVp530-MB (designated rHIVp530-MB196T and rHIVp530-MB196T). However, such recombinant clones also failed to replicate. It is unclear at this time as to how these recombinant HIV-1 virions failed to replicate;however, it is possible that other unidentified as yet critical AA substitutions are required for the replication of the DRV-selected mixture population. In this regard, all the recombinant HIV-1 clones with H219Q mutated back failed to replicate as we have previously published. Of note such recombinant HIV-1 virions such as HIV-1JRL75/H219/V390 and HIV-1NLL75R/H219/M390 were vitually totally replication-incompetent. The present results demonstrated the first successful in vitro selection of highly DRV-resistant HIV-1 variants and a new method for efficiently selecting drug-resistant HIV-1 variants in test tube when such variants are hardly generated in vitro and in vivo. The present data also suggest that DRV would not easily permit HIV-1 to develop significant resistance;however, HIV-1 can develop high levels of DRV-resistance with robust viral fitness comparable to the fitness of wild-type HIV-1 when super-infection with multi-PI-resistant HIV-1 variants exists and ensuing homologous recombination occurs.